Sei sulla pagina 1di 2

General Information

2018-00243 - [NEO] PhD Position / A Theme/Domain : Networks and


SINGULAR PERTURBATION APPROACH FOR Telecommunications
Scientific computing (BAP E)
MACHINE LEARNING IN MULTIOBJECTIVE Town/city : Avignon
MARKOV DECISION PROCESSES WITH Inria Center : CRI Sophia Antipolis -
LEXICOGRAPHIC OPTIMISATION Méditerranée
Starting date : 4/1/18
Contract type : Public service fixed-term contract Duration of contract : 3 years
Level of qualifications required : Graduate degree or equivalent
Deadline to apply : 3/31/18
Fonction : PhD Position

Contacts
About Inria Inria Team : NEO
Inria, the French National Institute for computer science and applied mathematics, Recruiter :
promotes “scientific excellence for technology transfer and society”. Graduates from Altman Eitan / eitan.altman@inria.fr
the world’s top universities, Inria's 2,700 employees rise to the challenges of digital
sciences. With its open, agile model, Inria is able to explore original approaches with
its partners in industry and academia and provide an efficient response to the
Conditions for
multidisciplinary and application challenges of the digital transformation. Inria is the application
source of many innovations that add value and create jobs.
Application file: Applications must be
submitted online on the Inria website.
About the research centre or Inria department Collecting applications by other
channels is not guaranteed.
The Inria Sophia Antipolis - Méditerranée center counts 37 research teams and 9
support departments. The center's staff (about 600 people including 400 Inria Defence Security :
employees) is composed of scientists of different nationalities (250 foreigners of 50 This position is likely to be situated in a
nationalities), engineers, technicians and administrators. 1/3 of the staff are civil restricted area (ZRR), as defined in
servants, the others are contractual. The majority of the research teams at the center
Decree No. 2011-1425 relating to the
are located in Sophia Antipolis and Nice in the Alpes-Maritimes. Six teams are based in
protection of national scientific and
Montpellier and a team is hosted by the computer science department of the
technical potential (PPST).Authorisation
University of Bologna in Italy. The Center is a member of the University and Institution
Community (ComUE) "Université Côte d'Azur (UCA)". to enter an area is granted by the
director of the unit, following a
favourable Ministerial decision, as
Context defined in the decree of 3 July 2012
This PhD position is within framework of the joint laboratory Inria - Nokia Bell Labs. relating to the PPST. An unfavourable
Ministerial decision in respect of a
position situated in a ZRR would result
Assignment in the cancellation of the appointment.

Recruitment Policy :
Multi-objective optimization is often handled by either searching for a Pareto optimal As part of its diversity policy, all Inria
solution or by formulation of the problem as a constrained optimization in which one positions are accessible to people with
optimizes one of the objectives and imposes constraints on all other objectives. We disabilities.
propose a new way to formulate multi-objectiive optimization as a lexicographic
problem. This is motivated by lexicographic optimization encountered in engineering,
such as max-min fairness definition in networking. Warning : you must enter your
For a problem with N objectives, a lexicographic optimization is defined in an N step e-mail address in order to save
recursive way, in which at step 1 we compute the set S(1) of optimal solutions of the your application to Inria.
optimization of the 1st objective, and at step i>1 we compute the set S(i) of solutions Applications must be submitted
of the optimization of the i-th objective restricted to the set S(i-1). online on the Inria website.
Within this research shall propose dynamic programming based solutions for the Processing of applications sent
problem of computing the sets S(i) and for obtaining lexicographic optimization of from other channels is not
multi-objective Markov decision problems. We shall then propose adaptive control guaranteed.
algorithms for combining learning and control in the case that the parameters of the
Markov decision problem are not known in advance. As optimization objective at each
steps we shall use the average cost criterion.

We shall extend this at the second and third years of the thesis to risk sensitive cost
criterion.

References:
E. Altman, Constrained Markov Decision Processes, Chapman and Hall/CRC, 1999.
Altman and A. Shwartz Adaptive Control of constrained Markov chains, IEEE Trans.
Auto. Control, Vol. 36 No. 4, pp. 454-462, 1991

Main activities
Write scientific papers for conferences and for journals, participate in patents

Skills
Technical skills and level required :

Languages : English

Relational skills : Works well in team

Benefits package
Subsidised catering service
Partially-reimbursed public transport
Social security
Paid leave
Flexible working hours
Sports facilities

Remuneration
Duration: 36 months
Location: Sophia Antipolis, France
Gross Salary per month: 1982€brut per month (year 1 & 2) and 2085€ brut/month
(year 3)